How do we inventory the use of generative AI in different areas of activity?

Over the last few weeks I keep circling back to the question of emerging practices with generative AI in higher education, as well as how they might complicate or problematise the assumptions being made in the still slightly limited policy discourse which has emerged around it. The problem is ‘generative AI’ is too nebulous a concept to adequately convey the uses to which the systems and tools being subsumed under it can be put within the sector. There’s an empirical question here – of an extremely descriptive and straight forward sort – which ought to be informing decision making. In its absence I worry that much of this discussion is being animated by the looming spectre of students passing off machine generated text as their own, whereas the opportunities and problems are vastly more complicated than this.